大视场波前编码成像系统中的图像复原

Image Restoration in a Wavefront Coding Imaging System with a Wide Field of View

大视场波前编码成像系统的点扩展函数(PSF)随视场变化敏感,传统上仅使用中心视场的PSF会使图像复原非常困难。基于PSF在各视场分区近似不变的思想,建立了大视场效应下的图像降质模型,而图像复原正是这一模型的逆过程;从而提出一种基于视场分区的维纳滤波图像复原方法,不同视场区域采用该区域中心的PSF进行复原,然后通过插值将各分区的复原图像融合为一体。该复原方法的模型更加精确,而且对不同视场区域可以选择不同的滤波器参数,具有更大的灵活性,测试结果表明该方法可以使复原质量得到明显改善。

The point spread function （PSF） is sensitive to the changes of view field in a wavefront coding system with a wide field of view （FOV）,which means that it is difficult to restore the image only using the central PSF as usually do. Image degradation model is established by partitioning the FOV into several blocks where inside each block PSF-invariance is assumed and image restoration is just the inverse of the degradation process. A Wiener filter restoration method based on FOV partition is proposed,which restores each block respectively using its corresponding PSF and combines them together by interpolation. The proposed method is based on a more accurate model and has flexibility in choosing filter parameters for different FOV blocks. Test results verify its superiority to the traditional method.